INDEX

Explanations

dates, names, titles

The neuron is detecting “By [Author Name]” bylines at the start of articles.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 pudieron

-1.36

)]);

-1.27

 querían

-1.26

}(),

-1.23

’

-1.17

大さじ

-1.16

ed

-1.13

 hogares

-1.13

 blurry

-1.11

PLICATION

-1.10

POSITIVE LOGITS

Frankreich

1.37

Were

1.34

eleste

1.29

Those

1.27

atação

1.27

êtements

1.24

chorus

1.23

droje

1.23

:");

1.20

fohl

1.20

Activations Density 0.124%